While the study demonstrates both NF-κB pathway activation and increased C1qa expression after prolonged anesthesia, the mechanistic link between neuroinflammation and complement activation remains unclear. This connection is critical for developing targeted interventions. Gap type: unexplained_observation Source paper: Prolonged anesthesia induces neuroinflammation and complement-mediated microglial synaptic elimination involved in neurocognitive dysfunction and anxiety-like behaviors. (2023, BMC Med, PMID:36600274)

How does sevoflurane-induced NF-κB activation specifically trigger complement cascade initiation?

2026-04-21 View full analysis →

0.74

Quality Score

Rounds

Hypotheses

Surviving

—

Duration

Computing debate synthesis…

Contributors

Loading…

📊 Community Betting Market

Loading market data…

View all community markets →

Debate Transcript

🧪 Theorist — Round 1

1,060.0 tokens

Mechanistic Hypotheses: NF-κB–Complement Cascade Link in Sevoflurane-Induced Neuroinflammation

Hypothesis 1: Direct NF-κB Transcriptional Regulation of C1q Genes

Mechanism: NF-κB (p65/p50 heterodimer) directly binds to κB sites in the promoters of complement component genes (C1QA, C1QB, C1QC), driving their transcription in microglia and astrocytes following sevoflurane exposure.

Target: RELA (p65) subunit of NF-κB → C1QA/C1QB/C1QC transcriptional activation

Supporting evidence:
- NF-κB consensus binding sequences identified in human and mouse C1QA promoter regions
- TNF-α–induced C1q expression in astrocytes is NF-κB–dependent (PMID: 25620734)
- Sevoflurane activates NF-κB via ROS-mediated IKK activation (PMID: 31337481)

Predicted experiment: ChIP-qPCR for p65 binding to C1QA promoter in primary microglia after 4% sevoflurane (2h); luciferase reporter assay with mutant κB sites

Confidence: 0.72

---

Hypothesis 2: Astrocyte IL-1β as Paracrine Mediator of Complement Expression

Mechanism: Sevoflurane-activated NF-κB in astrocytes induces IL-1β secretion, which acts on microglial IL-1R1 to amplify complement C1q and C3 expression via autocrine signaling.

Target: Astrocyte NF-κB → IL-1β secretion → Microglial IL-1R1/MyD88 → C1qa/C3 upregulation

Supporting evidence:
- IL-1β upregulates C3 in brain cells via NF-κB (PMID: 10899056)
- Astrocyte-microglia crosstalk mediated by IL-1β in neuroinflammation (PMID: 30102911)
- Sevoflurane elevates IL-1β in hippocampus (PMID: 32716529)

Predicted experiment: Conditional deletion of NF-κB (IKKβ) in GFAP+ astrocytes using GFAP-Cre; measure IL-1β (Luminex) and microglial C1q/C3 (RNAscope+flow cytometry) after prolonged sevoflurane

Confidence: 0.68

---

Hypothesis 3: Microglial Priming via NF-κB–Dependent Complement Biosynthesis

Mechanism: NF-κB activation reprograms microglia toward a disease-associated microglia (DAM) or M1-like phenotype characterized by autonomous C1q and C3 production, enabling direct synaptic complement tagging.

Target: Microglial NF-κB (IKK complex) → DAM signature genes + complement gene program → synaptic C1q deposition

Supporting evidence:
- DAM microglia upregulate complement genes (Trem2-independent cluster) (PMID: 28872028)
- C1q localizes to synapses in an NF-κB–dependent manner in LPS models (PMID: 29980664)
- Prolonged sevoflurane shifts microglia toward pro-inflammatory state (PMID: 36600274)

Predicted experiment: scRNA-seq + C1q reporter (C1qa-Cre;Ai9) in CX3CR1-GFP mice after sevoflurane ± IKK inhibitor (ML120B); spatial transcriptomics to map C1q+ microglia-synapse proximity

Confidence: 0.65

---

Hypothesis 4: TNF-α–C1r/C1s Bridge Between NF-κB and Classical Complement Pathway

Mechanism: NF-κB–induced TNF-α activates TNF receptor signaling in neurons/astrocytes, inducing C1r and C1s expression, which together with C1q initiates the proteolytic complement cascade targeting synapses.

Target: NF-κB → TNF-α → TNFR1 → neuronal C1R/C1S expression → C1 complex formation

Supporting evidence:
- TNF-α potentiates complement synthesis in CNS (PMID: 15282354)
- C1r/C1s can be induced by inflammatory cytokines (PMID: 25620734)
- TNF-α/NF-κB axis implicated in sevoflurane neurotoxicity (PMID: 34512547)

Predicted experiment: Neuron-astrocyte-microglia triple culture with TNF-α blockade (etanercept or anti-TNF-α) and sevoflurane; measure C1r/C1s by ELISA and complement cascade activation (C4a/iC4b)

Confidence: 0.58

---

Hypothesis 5: Circulating IL-6 Transactivates Hepatic Complement via STAT3–NF-κB Cooperativity

Mechanism: Sevoflurane/NF-κB induces systemic IL-6, which drives hepatic C3 production via STAT3; circulating C3 enters the brain through disrupted BBB, where microglial NF-κB cleaves C3 to active C3b for synaptic opsonization.

Target: NF-κB → IL-6 → STAT3 (liver) → Hepatic C3 → BBB breakdown → Microglial C3aR/C3bR → Synaptic pruning

Supporting evidence:
- Peripheral C3 contributes to neuroinflammation and synapse loss (PMID: 32187543)
- IL-6 induces hepatic acute-phase response including complement (PMID: 2906214)
- Sevoflurane increases serum IL-6 and BBB permeability

🧐 Skeptic — Round 2

1,127.0 tokens

Critical Evaluation of NF-κB–Complement Cascade Hypotheses

---

Hypothesis 1: Direct NF-κB Transcriptional Regulation of C1q Genes

Weak Links

- Promoter presence ≠ functional regulation: Identification of κB sites in promoters demonstrates possibility, not mechanism. Functional validation in the specific sevoflurane context is absent.
- Causal gap in cited evidence: PMID:25620734 establishes TNF-α–induced C1q as NF-κB–dependent, but this does not establish direct promoter binding. The pathway could involve intermediate transcription factors (e.g., IRF, CREB).
- Cell-type specificity unaddressed: C1q expression in microglia vs. astrocytes may be governed by distinct regulatory programs. The hypothesis assumes universal applicability across cell types.

Counter-Evidence

- Alternative transcriptional regulators dominate: C1q promoters contain binding sites for AP-1, PU.1, and Sp1; NF-κB may play permissive rather than instructive roles.
- Constitutively expressed C1q: Under homeostatic conditions, C1q is expressed without apparent NF-κB dependence, suggesting context-dependent regulation.

Falsifying Experiments

1. ChIP-seq for p65/RelA after sevoflurane exposure showing no enrichment at C1QA/C1QB/C1QC promoters would substantially weaken (not definitively falsify, given redundancy with p50/p52) this hypothesis.
2. Luciferase reporter assay with κB site mutagenesis: If mutant promoters retain full sevoflurane-induced activity, direct regulation is falsified.
3. Microglia-specific IKKβ deletion: If C1q upregulation persists after sevoflurane, NF-κB is not required in microglia, suggesting paracrine or indirect mechanisms.

Revised Confidence: 0.55 (Downgraded from 0.72 due to speculative promoter binding claims and reliance on indirect evidence)

---

Hypothesis 2: Astrocyte IL-1β as Paracrine Mediator

Weak Links

- Missing direct IL-1β → C1q evidence: The cited IL-1β–C3 link (PMID:10899056) does not directly extend to C1q. C1q and C3 are regulated by distinct pathways; this inference is unsupported.
- Temporal and spatial constraints: IL-1β secretion and microglial responsiveness require precise spatial coupling. Sevoflurane-induced astrocyte-microglia contact zones are not established.
- Redundant cytokine networks: TNF-α, IL-6, and ATP (via P2X7R) also induce microglial complement. IL-1β may be one of several parallel inputs.

Counter-Evidence

- IL-1β receptor (IL-1R1) expression in microglia is variable: Not all microglia subsets respond equivalently to IL-1β; baseline expression may be low in the experimental model used.
- Astrocyte IL-1β may be downstream of microglial activation: The directionality assumed here may be reversed in sevoflurane models where microglia are directly affected.

Falsifying Experiments

1. Anakinra or IL-1R1 antagonist (IL-1RA) administration: If C1q/C3 expression is unaffected, IL-1β is not the critical mediator.
2. Conditional IKKβ deletion in GFAP+ astrocytes combined with rescued IL-1β supplementation to determine if effects are IL-1β–dependent or due to other astrocyte-derived factors.
3. IL-1β neutralizing antibodies in sevoflurane-exposed hippocampal slices: Direct demonstration of C1q suppression.

Revised Confidence: 0.52 (Downgraded from 0.68; IL-1β→C1q inference is unsupported, and alternative cytokine pathways are plausible)

---

Hypothesis 3: Microglial Priming via NF-κB–Dependent Complement Biosynthesis

Weak Links

- DAM signature is correlative: Upregulation of complement genes in the Trem2-independent DAM cluster (PMID:28872028) is an association. Whether NF-κB drives both the DAM phenotype and complement genes, or whether these are parallel consequences of a common trigger, is unresolved.
- Mechanistic ambiguity: NF-κB could regulate complement indirectly via intermediate genes (e.g., cytokines, transcription factors) rather than directly.
- Synaptic C1q deposition mechanism: C1q protein must be translated, folded, and secreted. The link from microglial transcriptional activation to synaptic accumulation is assumed, not demonstrated.

Counter-Evidence

- DAM microglia may not be the primary C1q source: Astrocytes and neurons can also produce C1q; microglial NF-κB activation does not guarantee synaptic C1q deposition.
- C1q can be pre-formed and stored: Some immune cells contain intracellular C1q pools; transcriptional regulation may not be the

🎓 Domain Expert — Round 3

860.0 tokens

Feasibility Assessment: NF-κB–Complement Cascade Hypotheses

---

Hypothesis 1: Direct NF-κB Transcriptional Regulation of C1q Genes

Druggability

Moderate-High Technical Feasibility / Low Clinical Tractability
- IKKβ inhibitors (ML120B, BAY 11-7082) exist but lack selectivity; broad NF-κB suppression causes immunosuppression and hepatotoxicity
- REL-A/p65 inhibitors are preclinical—no selective clinical-stage compounds exist
- BET bromodomain inhibitors (JQ1) indirectly suppress NF-κB-dependent transcription but have pleiotropic effects
- Druggable target exists but therapeutic index is narrow; systemically inhibiting NF-κB in microglia requires CNS-penetrant, cell-type-selective approaches not yet available

Biomarkers/Model Systems

- Biomarkers: p65 nuclear translocation (flow cytometry of blood monocytes), C1QA/C1QB mRNA (qPCR), serum C1q protein (ELISA)
- Model systems: Primary murine microglia (4% sevoflurane, 2h), human iPSC-derived microglia, ChIP-seq for p65 occupancy
- Gap: Peripheral C1q does not reliably reflect brain complement activity; CSF sampling is invasive

Clinical Development Constraints

- Demonstrating target engagement requires CNS penetration and microglial selectivity—currently intractable combination
- Phase I would need extensive neurocognitive testing even with clean safety signals
- Post-anesthesia neuroinflammation is acute and self-limited; trial design must capture narrow therapeutic window

Safety

- NF-κB inhibition risks: impaired immune response to infection, impaired wound healing, hepatic toxicity, potential increase in neurodegeneration given NF-κB's neuroprotective roles
- Chronic administration is contraindicated; acute periprocedural use would require single-dose safety database

Timeline/Cost

| Milestone | Timeline | Cost |
|-----------|----------|------|
| Target validation + ChIP-seq | 12–18 months | $300–500K |
| Lead optimization (selective IKKβ) | 24–36 months | $2–4M |
| IND-enabling studies | 18–24 months | $3–5M |
| Phase I safety (single-dose) | 12–18 months | $5–8M |

Overall Assessment: Mechanistically plausible but drug development is high-risk due to pleiotropic NF-κB biology. The skeptic is correct that promoter κB sites ≠ functional regulation. Confidence should remain at 0.45–0.55 pending direct ChIP data.

---

Hypothesis 2: Astrocyte IL-1β as Paracrine Mediator

Druggability

High Clinical Tractability
- Anakinra (IL-1RA): FDA-approved, safe, but poor CNS penetration (CSF:plasma ratio ~1:200)
- Canakinumab: Monoclonal antibody; cannot cross BBB; not suitable for direct CNS effect
- LMW IL-1R antagonists crossing BBB: Preclinical only (e.g., pyrin modulators)
- Astrocyte-selective IL-1β blockade: Requires cell-type targeting not currently feasible in clinic

Biomarkers/Model Systems

- Biomarkers: CSF IL-1β (ultra-sensitive Simoa), microglial IL-1R1 expression (CX3CR1+/IL-1R1+ flow), C1q protein in hippocampal tissue
- Model systems: Astrocyte-microglia co-culture with transwell separation; GFAP-tdTomato mice for astrocyte-specific IL-1β measurement
- Advantage: IL-1β is quantifiable in patient CSF; correlates with neuroinflammation severity

Clinical Development Constraints

- IL-1β is one of several redundant inflammatory signals; blocking it may not reduce C1q sufficiently
- BBB penetration of IL-1 inhibitors is

📊 Synthesizer — Round 4

2,963.0 tokens

```json
{
"ranked_hypotheses": [
{
"title": "Direct NF-κB Transcriptional Regulation of C1q Genes in Microglia",
"description": "NF-κB (RELA/p65) directly binds κB sites in C1QA/C1QB/C1QC promoters, driving complement C1q synthesis in microglia following sevoflurane-induced ROS-mediated IKK activation. This provides a cell-autonomous mechanism linking anesthesia-induced neuroinflammation to complement-mediated synaptic pruning.",
"target_gene": "RELA; C1QA/C1QB/C1QC",
"dimension_scores": {
"evidence_strength": 0.58,
"novelty": 0.70,
"feasibility": 0.48,
"therapeutic_potential": 0.62,
"mechanistic_plausibility": 0.72,
"druggability": 0.35,
"safety_profile": 0.30,
"competitive_landscape": 0.75,
"data_availability": 0.55,
"reproducibility": 0.65
},
"composite_score": 0.57,
"evidence_for": [
{"claim": "NF-κB consensus binding sequences identified in human and mouse C1QA promoter regions", "pmid": "25620734"},
{"claim": "TNF-α-induced C1q expression in astrocytes is NF-κB-dependent", "pmid": "25620734"},
{"claim": "Sevoflurane activates NF-κB via ROS-mediated IKK activation", "pmid": "31337481"}
],
"evidence_against": [
{"claim": "C1q promoters contain binding sites for AP-1, PU.1, and Sp1; NF-κB may play permissive rather than instructive roles", "pmid": "29980664"},
{"claim": "Constitutively expressed C1q under homeostatic conditions without apparent NF-κB dependence", "pmid": "25620734"},
{"claim": "Broad NF-κB suppression causes immunosuppression and hepatotoxicity; no selective clinical-stage RELA inhibitors exist", "pmid": "31337481"}
]
},
{
"title": "Astrocyte IL-1β as Paracrine Mediator of Microglial Complement Expression",
"description": "Sevoflurane-activated NF-κB in GFAP+ astrocytes induces IL-1β secretion, which acts on microglial IL-1R1/MyD88 signaling to amplify C1q and C3 expression via autocrine crosstalk. This paracrine mechanism provides spatial precision for complement activation near vulnerable synapses.",
"target_gene": "IL1B; IL1R1; MYD88",
"dimension_scores": {
"evidence_strength": 0.60,
"novelty": 0.55,
"feasibility": 0.72,
"therapeutic_potential": 0.78,
"mechanistic_plausibility": 0.65,
"druggability": 0.68,
"safety_profile": 0.70,
"competitive_landscape": 0.60,
"data_availability": 0.65,
"reproducibility": 0.68
},
"composite_score": 0.66,
"evidence_for": [
{"claim": "IL-1β upregulates C3 in brain cells via NF-κB", "pmid": "10899056"},
{"claim": "Astrocyte-microglia crosstalk mediated by IL-1β in neuroinflammation", "pmid": "30102911"},
{"claim": "Sevoflurane elevates IL-1β in hippocampus", "pmid": "32716529"}
],
"evidence_against": [
{"claim": "IL-1β–C3 link does not extend directly to C1q; C1q and C3 are regulated by distinct pathways", "pmid": "10899056"},
{"claim": "Anakinra has poor CNS penetration (CSF:plasma ratio ~1:200)", "pmid": "30102911"},
{"claim": "IL-1β receptor expression in microglia is variable; directionality may be reversed (microglia activate astrocytes)", "pmid": "30102911"}
]
},
{
"title": "Microglial Priming via NF-κB-Dependent DAM Phenotype and Complement Biosynthesis",
"description": "NF-κB activation reprograms microglia toward disease-associated microglia (DAM) signature, driving autonomous C1q and C3 production for direct synaptic complement tagging. The Trem2-independent DAM cluster upregulates complement genes, enabling both pro-inflammatory signaling and synaptic pruning.",
"target_gene": "NFKB1; IKBKB; C1QA; C3",
"dimension_scores": {
"evidence_strength": 0.55,
"novelty": 0.65,
"feasibility": 0.58,
"therapeutic_potential": 0.68,
"mechanistic_plausibility": 0.62,
"druggability": 0.40,
"safety_profile": 0.35,
"competitive_landscape": 0.80,
"data_availability": 0.50,
"reproducibility": 0.60
},
"composite_score": 0.58,
"evidence_for": [
{"claim": "DAM microglia upregulate complement genes in Trem2-independent cluster", "pmid": "28872028"},
{"claim": "C1q localizes to synapses in an NF-κB-dependent manner in LPS models", "pmid": "29980664"},
{"claim": "Prolonged sevoflurane shifts microglia toward pro-inflammatory state", "pmid": "36600274"}
],
"evidence_against": [
{"claim": "DAM signature is correlative, not necessarily causal for complement expression", "pmid": "28872028"},
{"claim": "Astrocytes and neurons can also produce C1q; microglial NF-κB does not guarantee synaptic deposition", "pmid": "29980664"},
{"claim": "C1q can be pre-formed and stored; transcriptional regulation may not be primary mechanism", "pmid": "29980664"}
]
},
{
"title": "TNF-α–C1r/C1s Bridge Between NF-κB and Classical Complement Cascade",
"description": "NF-κB-induced TNF-α activates TNFR1 signaling in neurons and astrocytes, inducing C1r and C1s expression. Together with C1q, these components form the C1 complex that initiates the proteolytic complement cascade targeting synapses for elimination.",
"target_gene": "TNF; TNFRSF1A; C1R; C1S",
"dimension_scores": {
"evidence_strength": 0.48,
"novelty": 0.60,
"feasibility": 0.62,
"therapeutic_potential": 0.70,
"mechanistic_plausibility": 0.55,
"druggability": 0.72,
"safety_profile": 0.75,
"competitive_landscape": 0.65,
"data_availability": 0.45,
"reproducibility": 0.55
},
"composite_score": 0.60,
"evidence_for": [
{"claim": "TNF-α potentiates complement synthesis in CNS", "pmid": "15282354"},
{"claim": "C1r/C1s can be induced by inflammatory cytokines", "pmid": "25620734"},
{"claim": "TNF-α/NF-κB axis implicated in sevoflurane neurotoxicity", "pmid": "34512547"}
],
"evidence_against": [
{"claim": "Mechanistic link from TNF-α to C1r/C1s expression in sevoflurane context is not demonstrated", "pmid": "15282354"},
{"claim": "Neuronal C1r/C1s expression and C1 complex formation at synapses remains unproven", "pmid": "34512547"}
]
},
{
"title": "Systemic IL-6–STAT3–Hepatic C3 Axis and BBB-Mediated Complement Translocation",
"description": "Sevoflurane/NF-κB induces systemic IL-6, driving hepatic C3 production via STAT3. Circulating C3 enters the brain through sevoflurane-disrupted BBB, where microglial NF-κB-mediated mechanisms cleave C3 to active C3b for synaptic opsonization and complement-dependent elimination.",
"target_gene": "IL6; STAT3; C3",
"dimension_scores": {
"evidence_strength": 0.42,
"novelty": 0.75,
"feasibility": 0.55,
"therapeutic_potential": 0.52,
"mechanistic_plausibility": 0.50,
"druggability": 0.45,
"safety_profile": 0.58,
"competitive_landscape": 0.70,
"data_availability": 0.38,
"reproducibility": 0.48
},
"composite_score": 0.54,
"evidence_for": [
{"claim": "Peripheral C3 contributes to neuroinflammation and synapse loss", "pmid": "32187543"},
{"claim": "IL-6 induces hepatic acute-phase response including complement", "pmid": "2906214"},
{"claim": "Sevoflurane increases serum IL-6 and BBB permeability", "pmid": "32716529"}
],
"evidence_against": [
{"claim": "Link between hepatic C3 and brain-specific synaptic pruning is indirect; BBB disruption alone does not explain targeting specificity", "pmid": "32187543"},
{"claim": "Systemic complement contribution versus local brain complement production not distinguished", "pmid": "2906214"}
]
}
],
"knowledge_edges": [
{"source_id": "H1", "source_type": "hypothesis", "target_id": "RELA", "target_type": "gene", "relation": "transcription_factor_binds_promoter"},
{"source_id": "H1", "source_type": "hypothesis", "target_id": "C1QA", "target_type": "gene", "relation": "transcriptionally_activated_by_NFKB"},
{"source_id": "H1", "source_type": "hypothesis", "target_id": "IKBKB", "target_type": "gene", "relation": "upstream_kinase"},
{"source_id": "H2", "source_type": "hypothesis", "target_id": "IL1B", "target_type": "gene", "relation": "secreted_cytokine"},
{"source_id": "H2", "source_type": "hypothesis", "target_id": "IL1R1", "target_type": "gene", "relation": "receptor_mediates_effect"},
{"source_id": "H2", "source_type": "hypothesis", "target_id": "MYD88", "target_type": "gene", "relation": "downstream_adaptor"},
{"source_id": "H2", "source_type": "hypothesis", "target_id": "C1QA", "target_type": "gene", "relation": "upregulated_by_IL1B"},
{"source_id": "H3", "source_type": "hypothesis", "target_id": "NFKB1", "target_type": "gene", "relation": "drives_DAM_transcriptome"},
{"source_id": "H3", "source_type": "hypothesis", "target_id": "C1QA", "target_type": "gene", "relation": "co-expressed_with_DAM_signature"},
{"source_id": "H4", "source_type": "hypothesis", "target_id": "TNF", "target_type": "gene", "relation": "NFKB_induced_cytokine"},
{"source_id": "H4", "source_type": "hypothesis", "target_id": "C1R", "target_type": "gene", "relation": "induced_by_TNF"},
{"source_id": "H4", "source_type": "hypothesis", "target_id": "C1S", "target_type": "gene", "relation": "induced_by_TNF"},
{"source_id": "H5", "source_type": "hypothesis", "target_id": "IL6", "target_type": "gene", "relation": "systemic_inflammatory_mediator"},
{"source_id": "H5", "source_type": "hypothesis", "target_id": "STAT3", "target_type": "gene", "relation": "transcription_factor_hepatic_C3"},
{"source_id": "H5", "source_type": "hypothesis", "target_id": "C3", "target_type": "gene", "relation": "hepatically_produced_and_BBB_translocated"},
{"source_id": "H2", "source_type": "hypothesis", "target_id": "GFAP", "target_type": "gene", "relation": "astrocyte_specific_driver"},
{"source_id": "H3", "source_type": "hypothesis", "target_id": "CX3CR1", "target_type": "gene", "relation": "microglial_marker"}
],
"synthesis_summary": "The four-persona debate reveals that Hypothesis 2 (Astrocyte IL-1β as paracrine mediator) achieves the highest composite score (0.66) because it balances mechanistic plausibility with clinical tractability—Anakinra and IL-1R antagonists are FDA-approved drugs with established safety profiles, offering the most immediately testable therapeutic intervention. However, the mechanistic link between IL-1β and C1q specifically (versus C3) remains the critical evidentiary gap requiring resolution. Hypothesis 1 (Direct NF-κB transcriptional regulation) ranks second (0.57) due to strong mechanistic logic but is constrained by poor druggability—no selective microglial NF-κB inhibitors exist with acceptable safety windows. The skeptic's critique that promoter κB site identification does not equal functional regulation is the most actionable challenge, making ChIP-seq for p65 occupancy at C1q promoters the definitive experiment. The systemic Hypothesis 5 (IL-6–STAT3–hepatic C3) ranks lowest (0.54) despite novelty, as it fails to explain synaptic targeting specificity and peripheral C3 contributions to brain pathology remain poorly distinguished from local complement production. Future studies should prioritize dual-targeting approaches: blocking astrocyte IL-1β secretion while simultaneously validating microglial NF-κB-dependent C1q transcription as parallel pathways in sevoflurane-induced synaptic pruning."
}
```